🌱 STAYWELL

Early Detection of Academic Burnout Through Statistical Analysis

A Statistical Risk Assessment Platform Powered by Pure Mathematical Modeling

🎯 Overview

STAYWELL is an evidence-based early warning system designed to identify academic burnout risk among students using transparent statistical analysis. Unlike black-box machine learning approaches, our system employs pure mathematical modeling with full explainability and ethical compliance.

The Problem

Academic burnout affects student mental health, performance, and well-being. Early detection enables timely intervention and support.

Our Solution

A statistical risk assessment platform that analyzes measurable lifestyle indicators:

• 🛏️ Sleep patterns

• 📚 Study load

• 📱 Screen time

• 😰 Stress levels

• 📊 Academic engagement

✨ Key Features

📥 Multiple Input Methods

• CSV Upload: Batch analysis for entire cohorts

• Manual Entry: Individual student assessment with real-time validation

• Sample Data: Pre-loaded demonstration dataset

📊 Comprehensive Analytics

• Risk Overview: Distribution analysis and summary statistics

• Statistical Analysis: Correlations, distributions, descriptive statistics

• Individual Assessment: Per-student risk breakdown with factor contributions

• Scenario Simulation: What-if analysis for intervention planning

• Full Transparency: Complete methodology documentation

🎨 Visual Intelligence

• Risk distribution bar charts

• Time allocation visualizations

• Factor contribution breakdowns

• Correlation heatmaps

• Distribution histograms

• Color-coded risk indicators

🔍 Explainability First

• Transparent calculations

• Factor contribution analysis

• Statistical justifications

• Clear risk thresholds

• No black-box models

🔬 Statistical Methodology

Burnout Score Model

Weighted Linear Regression:

Burnout Score = Σ (Weight_i × Normalized_Factor_i)

Risk Factors & Weights

FactorWeightRationaleSleep Deficit30%Most critical for mental health and cognitive functionStress Level20%Direct indicator of psychological strainScreen Time20%Proxy for digital fatigue and time managementStudy Hours15%Excessive studying can lead to burnoutAttendance15%Low attendance indicates disengagement

Normalization Techniques

1. Sleep Deficit: max(0, (7 - sleep_hours) / 7)

2. Stress Level: stress_level / 5

3. Screen Time: screen_time / 10

4. Study Hours: study_hours / 10

5. Attendance: (100 - attendance) / 100

Risk Classification

• 🟢 Low Risk: Score < 0.30

• 🟡 Moderate Risk: 0.30 ≤ Score < 0.60

• 🔴 Elevated Risk: Score ≥ 0.60

Statistical Techniques

1. Descriptive Statistics: Mean, median, standard deviation, variance

2. Correlation Analysis: Pearson correlation coefficients

3. Percentile Analysis: Peer comparison using empirical distribution

4. Sensitivity Analysis: What-if scenario testing

5. Factor Decomposition: Proportional risk contribution

🚀 Quick Start

Prerequisites

Python 3.8+
pip package manager

Installation

# Clone the repository
git clone <repository-url>
cd staywell

# Install dependencies
pip install -r requirements.txt

Run Application

streamlit run app.py

Access at: http://localhost:8501

📁 Project Architecture

staywell/
│
├── app.py                          # Main Streamlit application
├── config.py                       # Model configuration & weights
├── requirements.txt                # Python dependencies
│
├── core/                           # Statistical Engine
│   ├── validation.py              # Input validation & constraints
│   ├── scoring_engine.py          # Burnout score calculation
│   ├── peer_engine.py             # Peer comparison statistics
│   └── statistical_analysis.py    # Statistical analysis functions
│
├── explainability/                 # Explainability Modules
│   ├── contribution.py            # Factor contribution analysis
│   └── what_if.py                 # Scenario simulation engine
│
├── ui/                            # User Interface
│   ├── dashboard.py               # Visualization components
│   └── theme.py                   # UI theme configuration
│
└── data/                          # Data Files
    ├── sample_students.csv        # Sample dataset
    └── test_invalid.csv           # Validation test data

📊 Data Format

CSV Structure

name,sleep_hours,study_hours,screen_time,stress_level,attendance
Alice Johnson,7.0,6.0,4.0,3,90
Bob Smith,8.0,5.0,3.0,2,95
Charlie Brown,6.0,7.0,5.0,4,82

Column Specifications

ColumnTypeRangeDescriptionnameString-Student identifier (optional)sleep_hoursFloat0-24Average daily sleep durationstudy_hoursFloat0-24Average daily study timescreen_timeFloat0-24Average daily recreational screen timestress_levelInteger1-5Self-reported stress (1=Low, 5=High)attendanceInteger0-100Class attendance percentage

⚠️ Important Constraint

24-Hour Rule: sleep_hours + study_hours + screen_time ≤ 24

Records violating this constraint are automatically filtered with notification.

🎯 Use Cases

For Educators

• Monitor cohort mental health trends

• Identify at-risk students early

• Plan targeted interventions

• Track effectiveness of support programs

For Counselors

• Prioritize student outreach

• Understand risk factor patterns

• Simulate intervention impacts

• Support evidence-based counseling

For Administrators

• Assess institutional wellness

• Allocate support resources

• Evaluate policy effectiveness

• Generate statistical reports

For Researchers

• Analyze burnout patterns

• Study factor correlations

• Test intervention strategies

• Validate statistical models

🔒 Ethical Framework

Core Principles

1. Transparency: All calculations are fully explainable

2. Privacy: Confidential data handling required

3. Support-Focused: Results guide help, not punishment

4. Human Oversight: Professional judgment essential

5. Not Diagnostic: Early warning only, not clinical diagnosis

Responsible Use

• ✅ Use for early identification and support

• ✅ Combine with professional counseling

• ✅ Maintain student privacy

• ✅ Focus on intervention, not labeling

• ❌ Do not use as sole diagnostic tool

• ❌ Do not stigmatize high-risk students

• ❌ Do not share results without consent

🛠️ Customization

Adjust Risk Weights

Edit config.py:

WEIGHTS = {
    "sleep": 0.30,      # Sleep deficit weight
    "stress": 0.20,     # Stress level weight
    "screen": 0.20,     # Screen time weight
    "study": 0.15,      # Study hours weight
    "attendance": 0.15  # Attendance weight
}

Modify Risk Thresholds

Edit core/scoring_engine.py:

def risk_label(score):
    if score < 0.3:     # Low risk threshold
        return "🟢 Low Risk"
    elif score < 0.6:   # Moderate risk threshold
        return "🟡 Moderate Risk"
    return "🔴 Elevated Risk"

Customize Visualizations

Edit ui/dashboard.py to modify charts, colors, and layouts.

📈 Domain-2 Compliance

✅ Pure Statistical Modeling - No machine learning black boxes
✅ Descriptive Analysis - Comprehensive statistical summaries
✅ Correlation Analysis - Factor relationship identification
✅ Regression-Based Scoring - Weighted linear model
✅ Clear Risk Thresholds - Statistically justified classifications
✅ Full Documentation - Complete methodology report
✅ Explainable Results - Transparent calculations
✅ Source Code Provided - Open and auditable

🧪 Testing

Test with Invalid Data

# Upload data/test_invalid.csv to see validation in action
# Contains records that violate 24-hour constraint

Manual Entry Testing

1. Select "Manual Entry" mode

2. Adjust time values to exceed 24 hours

3. Observe real-time validation feedback

4. Correct values to see risk assessment

📚 Documentation

• METHODOLOGY.md - Detailed statistical methodology

• AI_USAGE_SUMMARY.md - AI assistance documentation

• SIMPLE_UPDATE.md - Recent updates and changes

• Inline Comments - Code-level documentation

🤝 Contributing

This project is developed for educational purposes. Suggestions for improvements:

1. Additional risk factors

2. Enhanced visualizations

3. Export functionality

4. Multi-language support

5. Mobile responsiveness

📄 License

Educational project developed for Domain 2: Data & Statistical Modelling.

🙏 Acknowledgments

• Statistical methods based on psychological research

• Risk factors from academic burnout literature

• Ethical guidelines from mental health best practices

• UI/UX inspired by modern data platforms

📞 Support

For questions, issues, or suggestions:

• Review inline code documentation

• Check METHODOLOGY.md for detailed explanations

• Examine sample data for format examples

⚠️ Disclaimer

STAYWELL is an early warning system, not a diagnostic tool.

Always involve qualified mental health professionals for student support. This system provides statistical guidance to complement, not replace, professional judgment.

Built with ❤️ for Student Wellness

Empowering Early Intervention Through Statistical Intelligence